121

COPYRIGHT 2006 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED

Hip Displacement in
Cerebral Palsy
BY BRENDAN SOO, MBBS, JASON J. HOWARD, MD, FRCS(C), ROSLYN N. BOYD, PHD, MSC(PHYSIOTHERAPY),
SUSAN M. REID, MCLINEPI, ANNA LANIGAN, RN, RORY WOLFE, PHD,
DINAH REDDIHOUGH, MD, FRACP, FAFRM, AND H. KERR GRAHAM, MD, FRCS(ED), FRACS
Investigation performed at the Royal Childrens Hospital, Murdoch Childrens Research Institute,
University of Melbourne, Parkville, Victoria, Australia

Background: Hip displacement is considered to be common in children with cerebral palsy but the reported inci-
dence and the proposed risk factors vary widely. Knowledge regarding its overall incidence and associated risk fac-
tors can facilitate treatment of these children.
Methods: An inception cohort was generated from the Victorian Cerebral Palsy Register for the birth years 1990
through 1992, inclusive, and multiple data sources pertaining to the cohort were reviewed during 2004. Gross motor
function was assessed for each child and was graded according to the Gross Motor Function Classification System
(GMFCS), which is a valid, reliable, five-level ordinal grading system. Hip displacement, defined as a migration per-
centage of >30%, was measured on an anteroposterior radiograph of the pelvis with use of a reliable technique.
Results: A full data set was obtained for 323 (86%) of 374 children in the Register for the birth years 1990 through
1992. The mean duration of follow-up was eleven years and eight months. The incidence of hip displacement for the
entire birth cohort was 35%, and it showed a linear relationship with the level of gross motor function. The incidence
of hip displacement was 0% for children with GMFCS level I and 90% for those with GMFCS level V. Compared with
children with GMFCS level II, those with levels III, IV, and V had significantly higher relative risks of hip displacement
(2.7, 4.6, and 5.9, respectively).
Conclusions: Hip displacement is common in children with cerebral palsy, with an overall incidence of 35% found in
this study. The risk of hip displacement is directly related to gross motor function as graded with the Gross Motor
Function Classification System. This information may be important when assessing the risk of hip displacement for
an individual child who has cerebral palsy, for counseling parents, and in the design of screening programs and re-
source allocation.
Level of Evidence: Prognostic Level II. See Instructions to Authors for a complete description of levels of evidence.

C
erebral palsy, which has an incidence of approximately
two per 1000 live births, is the most common cause
of physical disability affecting children in developed
countries1. Although cerebral palsy is by definition a static en-
cephalopathy, the associated musculoskeletal pathology is usu-
ally progressive1-4. Hip displacement is thought to be common
in children with cerebral palsy and may progress from silent
subluxation to painful dislocation when left untreated2-5. Hip
displacement in children with cerebral palsy is usually attrib-
uted to spasticity and contracture of the hip adductors and
flexors as well as the medial hamstrings. This may result in
muscle imbalance as well as osseous deformity, including in-
creased femoral anteversion and acetabular dysplasia, which
further increases the risk of hip instability6-8. In younger chil-
dren with cerebral palsy, hip displacement is usually asymp-
tomatic, but the incidence of pain increases with the duration
of follow-up9. Pain and fixed deformity may contribute to dif-
ficulties with sitting, standing, walking, dressing, and perineal
hygiene. Unilateral hip dislocation is sometimes associated
with the development of pelvic obliquity and scoliosis10. The
reported incidence of hip displacement in children with cere-
bral palsy has ranged from 1% to 75% and has been correlated
with the severity of involvement and the ambulatory status11-13.
Children with mild involvement and who walk independently
are considered to have a low incidence of hip displacement,
whereas those with more severe involvement and who are un-
able to walk have the greatest risk of hip displacement12. The
variation in reported incidences reflects the selection bias
within previous studies, variable durations of follow-up, and
different definitions of hip displacement.
Many authors have reported that early surgical interven-
tion in patients with spastic hip disease leads to better long-
term outcomes and decreases the risk of treatment failures14.
Some centers have established surveillance programs to iden-

THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006
tify the hip-at-risk in patients with cerebral palsy, so that
treatment may be implemented prior to the onset of symp-
tomatic dislocation15,16. Effective programs must be based on
an understanding of the incidence and natural history of hip
displacement in cerebral palsy as it relates to disease severity.
Familiarity with incidence rates may also help clinicians to be
more aware of the relative risk of hip disease when caring for a
child with cerebral palsy.
The hypothesis central to this study was that the inci-
dence of hip displacement in children with cerebral palsy is di-
rectly related to the level of gross motor function. Incidence is
defined as the proportion of new cases with a specific condi-
tion in the population at risk during a specified time period17.
Thus, to determine the precise incidence of hip displacement,
the population at risk must first be identified. We used the Ce-
rebral Palsy Register for the State of Victoria, Australia, to
identify a large, population-based sample of children with ce-
rebral palsy.
Cerebral palsy may be classified according to motor type,
topographical distribution, and functional severity. The motor
types can be described as spastic, dystonic, mixed, ataxic, or hy-
potonic, and the definitions of each type have recently been
updated1,18-20. The most common topographical distributions
are spastic hemiplegia, spastic diplegia, and spastic quadri-
plegia1,21. Spastic hemiplegia is characterized by unilateral in-
volvement, and spastic diplegia and spastic quadriplegia are
characterized by bilateral involvement. However, neither the de-
scriptors nor the definitions of the terms for motor type and to-
pographical distribution have been agreed upon. Furthermore,
it is known that classifications based on motor type and topo-
graphical distribution may not be reliable1,22.
The most useful development in the classification of
cerebral palsy in recent years has been the creation of the
Gross Motor Function Classification System (GMFCS)23. The
GMFCS is a five-level ordinal grading system based on the
assessment of self-initiated movement with emphasis on
function with regard to sitting and walking (Fig. 1). Distinc-
tions between levels are based on functional limitations, the
need for walking aids or wheeled mobility equipment, and
quality of movement according to age. Children, six to
twelve years of age, who have GMFCS level-I function have a
nearly normal level of gross motor function. Children with
level-II function walk independently but have limitations in
activities such as running or jumping. Those with level-III
function require assistive devices to walk and use a wheel-
chair for longer distances. Those with level-IV function have
the ability to stand for transfers, have minimal walking abil-
ity, and depend mainly on a wheelchair for mobility. Chil-
dren with level-V function lack head control; cannot sit
independently, stand, or walk; and are dependent for all as-
pects of care. Unlike the classifications of motor type and to-
pographical distribution, the GMFCS has been confirmed in
a number of studies to be a valid, reliable, stable, and clini-
cally relevant method for the classification and prediction of
motor function of children with cerebral palsy between the
ages of two and twelve years24.
122
HIP DISPLACEMENT IN C E RE B R A L P A L S Y
Materials and Methods
sing the Cerebral Palsy Register for the State of Victoria,
U Australia, we identified a population-based sample of 374
children, born between January 1990 and December 1992, who
had a confirmed diagnosis of cerebral palsy; multiple data
sources pertaining to the cohort were reviewed during 2004.
These three birth years were chosen to provide a large popula-
tion-based sample of children with cerebral palsy of sufficient
maturity to allow an accurate determination of the incidence of
hip displacement and gross motor function, according to the
GMFCS. Detailed information about the Victorian Cerebral
Palsy Register and the composition of this cohort has been pub-
lished elsewhere25. The Register contains prospectively gathered
demographic and clinical information about children, born in a
specific area, who had a confirmed diagnosis of cerebral palsy
according to accepted definitions1,20. Details concerning children
with a suspected diagnosis of cerebral palsy are forwarded to
the Director of the Child Development and Rehabilitation
Program, who is also the Director of the Victorian Cerebral
Palsy Register. The information is reviewed, and additional ex-
aminations or investigations can be requested. The evidence is
weighed, and the child may be registered as having cerebral
palsy. Registration is based on a clinical diagnosis of a cerebral
palsy syndrome. The results of brain imaging, either magnetic
resonance imaging or computed tomography, are available for
the majority of patients but are not mandatory for registration
purposes. However, if imaging or another investigation reveals
an alternative diagnosis, the diagnosis of cerebral palsy is re-
viewed and the child may be removed from the Register. The
Register is frequently reviewed and updated in light of informa-
tion from multiple sources.
The Register also includes classification of the cerebral
palsy according to motor type; topographical distribution;
and, more recently, gross motor function according to the
Gross Motor Function Classification System23,25. The Register
is maintained in accordance with state and institutional regu-
lations with respect to privacy legislation and ethical consider-
ations. Our institutional review board approved this study.
The inclusion criteria for the study were (1) a birth date
between January 1990 and December 1992, inclusive; (2) reg-
istration of the child in the Victorian Cerebral Palsy Register
prior to the commencement of this study; (3) confirmation of
a diagnosis of cerebral palsy by at least two independent physi-
ciansi.e., the physician who referred the child for registra-
tion and a senior clinician from another hospital department,
including Pediatric Medicine, Pediatric Neurology, or Clinical
Genetics; (4) adequate clinical records to allow determination
of the functional level according to the Gross Motor Function
Classification System for children between the ages of six and
twelve years; (5) survival to the age of six years; and (6) at least
one good-quality anteroposterior pelvic radiograph made be-
tween the age of six years and the time of final follow-up. The
exclusion criteria were (1) an absence of one or more inclu-
sion criteria and (2) loss to follow-up because of an inability
to contact the family.
Children who are registered as having cerebral palsy are
 123
THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG HIP DISPLACEMENT IN C E RE B R A L P A L S Y
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006

routinely assessed in a Hip Surveillance Clinic dedicated to the
early diagnosis and management of hip displacement in chil-
dren with cerebral palsy. A standardized clinical and radio-
graphic evaluation of hip development is undertaken in this
clinic, and systematic follow-up according to the perceived risk
of hip displacement is offered. This follow-up ranges from a
single pelvic radiograph after the age of six years for children
with spastic hemiplegia to a radiograph made every six to
twelve months for all children with hypertonia who are unable
to walk. When hip displacement is detected, appropriate inter-

Fig. 1
The Gross Motor Function Classification System (GMFCS) for children with cerebral palsy between the ages of
six and twelve years, which was developed by Palisano et al.23. (Reprinted, with permission, from: Graham HK.
Classifying cerebral palsy. J Pediatr Orthop. 2005;25:128.)

THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006
Fig. 2
Gross motor function of children with unilateral cerebral palsy (spastic hemiplegia). The majority
of these children function at Gross Motor Function Classification System (GMFCS) level I or II.
vention is offered according to clinical indications and the fam-
ilys wishes15,16.
Each childs level of gross motor function was classified
according to the Gross Motor Function Classification System,
with use of the descriptors for children between six and twelve
years of age23. The five GMFCS levels for children in this age
range are illustrated in Figure 1, and a more complete descrip-
tion of the GMFCS system can be found elsewhere23,26. The
GMFCS is simple and easy to learn, and it does not require
special training or equipment. It takes an experienced ob-
server about five minutes to assign a GMFCS level. Multiple
records were checked for agreement concerning the level, and
discrepancies were resolved by discussion with the childs
physiotherapist and developmental pediatrician. Children
without an assigned GMFCS level were classified by at least
two of the authors of this study after discussion with the
childs community and hospital-based physiotherapist.
The primary outcome measure in this study was the
presence of hip displacement, defined as a migration percent-
age of >30% in one or both hips. The migration percentage is
measured by drawing a Hilgenreiner line and Perkins line on
an anteroposterior radiograph of the hips27. The amount of
ossified femoral head that lies lateral to the Perkins line is
measured, divided by the width of the femoral head, and mul-
tiplied by 100, to be expressed as a percentage28,29. The migra-
tion percentage is a linear measure of hip displacement and is
the most valid, reliable, and useful measure of hip displace-
ment in children with cerebral palsy27-32. The migration per-
centage was measured on all available hip radiographs with
use of previously described methodology, which has been
shown to be reliable28. Reliability can be improved to accept-
able levels by special training of the observers28,29. We could not
study the natural history of hip displacement in our population
because of a high intervention rate. However, our management
policy is to not intervene until the migration percentage exceeds
30%. Such hips are unequivocally abnormal, although not all
require or receive intervention27,30,32,33.
The majority of children with severe cerebral palsy had
multiple radiographs made during childhood. Hip displace-
124
HIP DISPLACEMENT IN C E RE B R A L P A L S Y
ment was recorded when one or both hips had a migration
percentage of >30%. Hip dislocation (a migration percentage
of >100%) and all surgical procedures on the hips were re-
corded as well. Operations for hip displacement were classi-
fied as preventative (adductor or psoas release), reconstructive
(femoral or pelvic osteotomy), and salvage (excision arthro-
plasty, interposition arthroplasty, replacement arthroplasty, or
valgus proximal femoral osteotomy)31.
The children were stratified by motor type, topographi-
cal distribution, and GMFCS level, and the incidence of hip
displacement within those groups was determined. The risk of
hip displacement was calculated as the relative risk, or risk ra-
tio, for each motor type and topographical distribution as
compared with spastic diplegia and for each GMFCS level as
compared with level II. The confidence intervals were deter-
mined for all of the risk ratios.
Results
here were 374 children in the Victorian Cerebral Palsy
TRegister with a birth date between January 1990 and De-
cember 1992. Nine patients, with no recorded evidence of hip
displacement, died before the age of six years and were ex-
cluded from the study. An additional twenty-six patients were
excluded because of inadequate clinical or radiographic
records, and sixteen more were lost to follow-up. Complete
records were thus available for 323 children (86% of the origi-
nal birth cohort). The incidence of cerebral palsy during the
study period was 1.98 per 1000 live births, which is similar to
the incidence determined from other large cerebral palsy reg-
isters and confirms a high ascertainment rate.
A total of 279 children, 86% of the cohort, had a spastic
motor type, with approximately one-third each having spastic
hemiplegia, spastic diplegia, and spastic quadriplegia. Five chil-
dren had dystonia, twenty-one had a mixed pattern (spastic and
dystonic), nine had ataxia, and nine had hypotonia. Thirty-five
percent of the children had level-I gross motor function accord-
ing to the GMFCS. The percentages of patients with levels II, III,
IV, and V function were 16.4%, 14.2%, 16.1%, and 18%, re-
spectively. The majority of children with spastic hemiplegia had
 125
THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG HIP DISPLACEMENT IN C E RE B R A L P A L S Y
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006

Fig. 3
The gross motor function of children with bilateral cerebral palsy (spastic diplegia or spastic
quadriplegia) is distributed across all Gross Motor Function Classification System (GMFCS) lev-
els (I through V). The majority of children with spastic diplegia (white part of bar) function at GM-
FCS level I, II, or III, and the majority of children with spastic quadriplegia (black part of bar)
function at GMFCS level III, IV, or V. However, there is a continuum of involvement, and the sepa-
ration of children into the diplegic and quadriplegic groups is somewhat arbitrary, in terms of
gross motor function.

GMFCS level-I or II function (Fig. 2), whereas those with bilat-
eral cerebral palsy (spastic diplegia and spastic quadriplegia)
were distributed across all GMFCS levels (Fig. 3).
The mean duration of follow-up (and standard devia-
tion) was 11.7 0.17 years (range, six years and six months to
fourteen years). The mean number of radiographs available
for review was 1.8 0.4 (range, one to five) for children with
spastic hemiplegia, 2.3 1.4 (range, one to seven) for children
with spastic diplegia, and 6.8 2.8 (range, two to twenty-
four) for children with spastic quadriplegia.
One hundred and fourteen (35.3%) of the children had
hip displacement, defined as a migration percentage of >30%.
Twenty-one of the children with hip displacement (6.5% of
the cohort) had dislocation of one or both hips, defined as a

Fig. 4
Incidence of hip displacement (a migration percentage of >30%) according to the Gross Mo-
tor Function Classification System (GMFCS) level.
 126
THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG HIP DISPLACEMENT IN C E RE B R A L P A L S Y
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006

TABLE I Relative Risk of Hip Displacement* at Each GMFCS Level as Compared with Level II

GMFCS Level Relative Risk P Value 95% Confidence Interval

I 0
II 1
III 2.7 0.01 1.3 to 5.7
IV 4.6 <0.001 2.4 to 8.9
V 5.9 <0.001 3.1 to 11.3

*Hip displacement was defined as a migration percentage of >30%.

migration percentage of >100% at some point. One hundred
of the 114 children with hip displacement had one or more
operations to correct or prevent further displacement. Forty-
two children had preventative surgery, sixty-six had a combi-
nation of preventative and reconstructive surgery, and three
children had salvage surgery, for one or both hips31. (Some
children required combinations of surgery [for example, re-
constructive surgery following failed preventative surgery or
salvage surgery after failed reconstructive surgery].)
The lowest incidence of hip displacement was found in
the children with GMFCS level-I function and the highest in-
cidence, in those with level-V. None of the 114 children with
level-I function had hip displacement or hip surgery. In com-
parison, 90% of the children with level-V function had hip
displacement. The relationship between the GMFCS level and
hip displacement was essentially linear, with the incidence of
hip disease increasing in direct proportion to functional sever-
ity, as demonstrated in Figure 4. This relationship was also
shown by calculations of the relative risks of hip displacement
associated with the functional levels (Table I). Relative risk is
determined by a ratio of proportions. Since the incidence of
hip displacement in the children with level-I function was
zero, level II was chosen as the denominator in these calcula-
tions. When compared with level II, levels III, IV, and V
showed significantly different, and increasingly higher, risk ra-
tios for hip displacement (2.7, 4.6, and 5.9, respectively). Dis-
location was found in six (12%) of fifty-two children with
level-IV function and fifteen (26%) of fifty-eight children with
level-V function.
The majority of children with hip displacement had a
spastic motor type of cerebral palsy. Not only is this consid-
ered to be the highest risk category, but it was by far the largest
subgroup, representing 86% of the cohort. Two of the five pa-
tients classified as having dystonic cerebral palsy and seven of
the twenty-one classified as having a mixed motor type had
hip displacement. Thus, the incidence of hip displacement was
similar for all hypertonic motor types (spastic, dystonic, and
mixed). The only motor type with a 0% incidence of hip dis-
placement was the ataxic group. Four of the nine children in
the hypotonic group had hip displacement, which was associ-
ated with coxa valga and acetabular dysplasia rather than ad-
ductor spasticity and contracture. The timing and type of hip
displacement appeared to vary with the type of cerebral palsy,
but the small numbers in some groups precluded a statistical
analysis.
The incidences and relative risks for hip displacement
according to motor type and topography are shown in Tables
II and III.

TABLE II Incidence of Hip Displacement* According to Motor Type and Topographical Distribution

Hip Instability
Motor Type/Topographical Pattern Present Not Present Total
Hypertonic
Spastic
Spastic hemiplegia 1 (1.0%) 97 (99.0%) 98
Spastic diplegia 15 (19.2%) 63 (80.8%) 78
Spastic quadriplegia 85 (82.5%) 18 (17.5%) 103
Dystonic 2 (40.0%) 3 (60.0%) 5
Mixed 7 (33.3%) 14 (66.7%) 21
Ataxic 0 (0.0%) 9 (100.0%) 9
Hypotonic 4 (44.4%) 5 (55.6%) 9
Total 114 (35.3%) 209 (64.7%) 323

*Hip displacement was defined as a migration percentage of >30%. The values are given as the number of children with the percentage in
parentheses.
 127
THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006
TABLE III Relative Risk of Hip Displacement* for each Motor Type and Topographical Distribution as Compared with Spastic
Diplegia
Motor Type/Topography Relative Risk
Spastic hemiplegia 0.11
Spastic diplegia 1
Spastic quadriplegia 4.3
Dystonic 2.1
Mixed 1.7
Ataxic 0
Hypotonic 2.3
*Hip displacement was defined as a migration percentage of >30%.
Discussion
quinus and hip displacement are the two most common
E musculoskeletal deformities found in children with cerebral
palsy7,21. Estimates of the incidence of hip displacement in cere-
bral palsy have ranged from 2% to 75%, with the outcome of
the hip instability being significantly influenced by the age at
the time of the diagnosis and the intervention for this common
problem12,15,21,31,32. In patients with severe cerebral palsy, hip dis-
placement is often diagnosed late as a result of the silent nature
of the displacement, communication difficulties, and the in-
creased attention paid to other important issues such as feeding
difficulties and seizure management. Identifying the subset of
patients with an increased risk of hip displacement is essential
in planning surveillance programs and early intervention15,16.
Many authors have reported incidences and incidence rates of
hip displacement in an attempt to quantify the risk and to guide
screening protocols. To date, however, almost all studies have
relied on retrospective analyses based on referred cohorts,
which are not representative of the cerebral palsy population as
a whole. In addition, most of these studies have lacked a valid
and reliable method to measure function, making it difficult to
relate the incidence of hip displacement to the varying degrees
of disease severity found in these children12,13,32.
To our knowledge, the only previously reported popula-
tion-based study of hip disease in patients with cerebral palsy
was by Scrutton et al., who prospectively studied a large co-
hort of children, born with cerebral palsy, who ranged in age
from eighteen months to five years old33. The migration per-
centage measured in the children in that study had diverged
from normal by the age of eighteen months, and, by the age of
five years, 36% of the children needed some form of treat-
ment. However, follow-up to the age of five years is insuffi-
cient for a complete ascertainment of the incidence of hip
disease. This is made evident by the low (52%) incidence of
hip displacement in children with spastic quadriplegia re-
ported in their study as compared with incidence rates of 60%
to 80% reported in previous studies6,12,14,32. Their study showed
that the incidence of hip displacement is higher in children
with hypertonic motor disorders and is related to both the se-
HIP DISPLACEMENT IN C E RE B R A L P A L S Y
P Value 95% Confidence Interval
0.002 0.03 to 0.45
<0.001 2.7 to 6.8
0.22 0.6 to 6.7
0.15 0.8 to 3.7
0.06 1.0 to 5.5
verity of involvement and the ambulatory status, a claim sup-
ported by most previous studies of hip disease in individuals
with cerebral palsy12,13,34-37.
Using a cerebral palsy register with a high level of case
ascertainment, we identified a three-year birth cohort to gen-
erate a highly representative population sample of children
with cerebral palsy. All etiologies and levels of severity were
represented in this sample. We acknowledge that it would be
necessary to follow all children until skeletal maturity to
quantify the absolute incidence of hip displacement since a
few hips displace for the first time during the growth spurt as-
sociated with puberty or when scoliosis develops. However, we
think that this is a relatively small number of hips; most hips
that are going to dislocate do so by the age of ten years32. Our
choice of when to censor the study population was influenced
by the fact that the GMFCS has not yet been validated in older
populations. In addition, at our center, increasing numbers of
children with mild cerebral palsy are lost to follow-up after the
age of ten years. The distribution of children throughout all
GMFCS levels (Figs. 2 and 3) confirms the wide range of func-
tion and disability in a typical population sample of children
with cerebral palsy. Children with GMFCS level-I function are
completely independent, do not use aids, usually have a mild
spastic hemiplegia, and do not have hip displacement. Their
relatively normal muscle tone and high level of activity seem
to be protective against hip disease. Only one child with spas-
tic hemiplegia, which was Type IV according to the classifica-
tion described by Winters et al.38, required reconstructive
surgery for a subluxated hip. In contrast, children with GM-
FCS level-V function lack head control, have no sitting bal-
ance, have no independent mobility, and usually have severe
spastic dystonia in a quadriplegic pattern. Such children have
a 90% incidence of hip displacement (Fig. 4). The combina-
tion of severe bilateral spasticity and immobility may be re-
sponsible for the very high incidence of hip displacement in
these children. The incidences of hip displacement associated
with GMFCS levels II, III, and IV are intermediate between
these two extremes. Thus, there was a steadily increasing risk
of hip displacement with increasing functional severity in this

THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006
cohort. A nearly straight-line relationship between the rate of
hip displacement and the GMFCS level was apparent (Fig. 4).
This suggests that the GMFCS provides a reliable representa-
tion of motor function and further substantiates its use in the
classification of the severity of cerebral palsy39,40.
Understanding the relationship between the GMFCS and
the traditional motor-type and topographical classifications is
important to appreciate the utility of the GMFCS and to pro-
vide a reference point for its use40. The descriptors for the mo-
tor types of cerebral palsy have not been universally agreed on;
new definitions are under development, and the definitions
may be difficult to apply in a reliable manner1,18-20,22,39,40. Dystonia
may develop for the first time in late childhood or adolescence
in a child previously described as having a spastic motor
type41,42. The key finding of our study was that all hypertonic
motor types (spastic, dystonic, and mixed) were associated
with a similar incidence of hip displacement. The overall inci-
dence of hip displacement associated with a spastic motor type
of cerebral palsy was 36% compared with 40% for dystonia and
33% for the mixed motor type. With the numbers available,
these differences were not significant (Tables II and III). From
the point of view of predicting hip displacement, knowing the
childs GMFCS level is more important than defining whether
the muscle tone is spastic, dystonic, or mixed. The incidence of
hip displacement in children with hypotonia was similar to
the incidence in children with hypertonia, but the clinical and
radiographic features of the displacement were different. In
children with hypotonia, contractures did not develop until a
fixed hip dislocation occurred, but all of the children had se-
vere coxa valga and some had acetabular dysplasia. Ataxia was
the only motor type not associated with hip displacement.
The study by Howard et al. was one of the first to demon-
strate the usefulness of the traditional motor type/topographical
classification as a tool for predicting hip displacement in chil-
dren with cerebral palsy12. The reason for the strong relationship
between hip displacement and the classification of cerebral
palsy as spastic hemiplegia, spastic diplegia, or spastic quadri-
plegia is that these classifications are surrogates for gross motor
function. However, topographical classification of cerebral palsy
can be difficult, especially in terms of differentiation of unilat-
eral involvement from bilateral involvement with severe
asymmetry1. In addition, the classification of cerebral palsy with
bilateral involvement as diplegia or quadriplegia is somewhat
arbitrary and unsatisfactory1,39,40. Almost all children with spas-
tic hemiplegia walk independently; most children with spastic
diplegia walk, although many need assistance; and most chil-
dren with spastic quadriplegia do not walk. This simple classifi-
cation is not as useful as the GMFCS, which allows classification
of children with diplegia and quadriplegia according to their
functional level in a valid and reliable manner40. There are im-
portant differences in both the incidence and the severity of hip
displacement in children who have spastic quadriplegia at GM-
FCS level IV as compared with children with GMFCS level V.
However, we suggest that a complete description of the cerebral
palsy phenotype should include the motor type, topographical
distribution, and gross motor function19,20,26,39.
128
HIP DISPLACEMENT IN C E RE B R A L P A L S Y
The majority of previous studies support a migration
percentage of >30% as a suitable definition of hip displace-
ment. Hips with such displacement are clearly abnormal, and
the displacement is liable to progress in children with severe
involvement27,30,33. We chose this threshold for two additional
reasons. No child had intervention to address the hip displace-
ment until the migration percentage exceeded 30%, and the
majority (100) of the 114 who did reach that migration per-
centage required at least one operative procedure to correct
the hip displacement.
The high incidence rates and comparative ease of diag-
nosis of hip disease in children with cerebral palsy make a
compelling case for early detection. The incidence of hip dis-
placement within this representative population of children
with cerebral palsy was 35% compared with a 0.2% incidence
of developmental dysplasia of the hip within the general pop-
ulation. Hip surveillance programs for children with cerebral
palsy have shown some promise in both England and Aus-
tralia15,16. We agree with Scrutton and Baird15, who suggested
that all children with bilateral cerebral palsy (spastic diplegia
or quadriplegia) should have a first hip radiograph by the age
of thirty months. The frequency of follow-up radiographs
can be guided by the GMFCS level. Because children with
GMFCS level-IV or V function have such a high incidence of
hip displacement, we think that making a radiograph every
twelve to twenty-four months until skeletal maturity would
be clinically and economically justifiable. More frequent ra-
diographs are required for children who have a rapid in-
crease in the migration percentage or who have a migration
percentage of >30% and those in whom scoliosis or pelvic
obliquity develops. The incidence of hip displacement in
children with spastic hemiplegia is very low, and the dis-
placement develops later than it does in children with bilat-
eral cerebral palsy. We recommend a single hip radiograph,
made at the age of six to eight years, for children with spastic
hemiplegia. Only children with proximal involvement (the
Type-IV gait pattern described by Winters et al.38) require ad-
ditional radiographs.
The strengths of this study were that it was population-
based, with identification of cases through the use of a
statewide register of children with cerebral palsy, and it incor-
porated a valid, reliable measure of gross motor function (the
GMFCS). The weaknesses of this study were the variable tim-
ing of clinical and radiographic assessment of hip disease
within the cohort as well as the marked variability in age at the
times of detection and intervention.
To our knowledge, this is the first population-based
study of children with cerebral palsy that evaluated the inci-
dence of hip displacement according to disease severity as de-
fined by the GMFCS. This study confirms the high incidence
of hip displacement in children with cerebral palsy and the di-
rect relationship with functional level. These findings provide
surgeons with an appreciation of the relative risk of hip dis-
placement for the individual child with cerebral palsy, which
could be used as a guide to determine the resources required
for population screening and a targeted surgical program. 
 129
THE JOUR NAL OF BONE & JOINT SURGER Y JBJS.ORG HIP DISPLACEMENT IN C E RE B R A L P A L S Y
VO L U M E 88-A N U M B E R 1 J A N U A R Y 2006

Brendan Soo, MBBS Monash University, Alfred Hospital, Commercial Road, Prahran, Victo-
Jason J. Howard, MD, FRCS(C) ria 3004, Australia
Roslyn N. Boyd, PhD, MSc(Physiotherapy)
Susan M. Reid, MClinEpi In support of their research for or preparation of this manuscript, one
Anna Lanigan, RN or more of the authors received a Murdoch Childrens Research Institute
Dinah Reddihough, MD, FRACP, FAFRM Theme Grant. None of the authors received payments or other benefits
H. Kerr Graham, MD, FRCS(Ed), FRACS or a commitment or agreement to provide such benefits from a com-
Departments of Orthopaedic Surgery (B.S., J.J.H., R.N.B., mercial entity. No commercial entity paid or directed, or agreed to pay
and H.K.G.) and Child Development and Rehabilitation (S.M.R., or direct, any benefits to any research fund, foundation, educational in-
A.L., and D.R.), Royal Childrens Hospital, Flemington Road, stitution, or other charitable or nonprofit organization with which the
Parkville, Victoria 3052, Australia. E-mail address for H.K. Graham: authors are affiliated or associated.
kerr.graham@rch.org.au

Rory Wolfe, PhD doi:10.2106/JBJS.E.00071

References
1. Stanley F, Blair E, Alberman E. Cerebral palsies: epidemiology and causal path- 23. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Develop-
ways. London: Mac Keith; 2000. ment and reliability of a system to classify gross motor function in children with
cerebral palsy. Dev Med Child Neurol. 1997;39:214-23.
2. Graham HK. Painful hip dislocation in cerebral palsy. Lancet. 2002;359:907-8.
24. Wood E, Rosenbaum P. The gross motor function classification system for ce-
3. Graham HK. Botulinum toxin type A management of spasticity in the context of
rebral palsy: a study of reliability and stability over time. Dev Med Child Neurol.
orthopaedic surgery for children with spastic cerebral palsy. Eur J Neurol. 2001;8
2000;42:292-6.
Suppl 5:30-9.
25. Howard J, Soo B, Graham HK, Boyd RN, Reid S, Lanigan A, Wolfe R, Reddi-
4. Graham HK. Mechanisms of deformity. In: Scrutton D, Damiano D, Mayston M,
hough DS. Cerebral palsy in Victoria: motor types, topography and gross motor
editors. Management of the motor disorders of children with cerebral palsy. 2nd
function. J Paediatr Child Health. 2005;41:479-83.
ed. London: Mac Keith; 2004. p 105-29.
26. Graham HK. Classifying cerebral palsy. J Pediatr Orthop. 2005;25:127-8.
5. Bagg MR, Farber J, Miller F. Long-term follow-up of hip subluxation in cerebral
palsy patients. J Pediatr Orthop. 1993;13:32-6. 27. Reimers J. The stability of the hip in children. A radiological study of the results
of muscle surgery in cerebral palsy. Acta Orthop Scand Suppl. 1980;184:1-100.
6. Noonan KJ, Walker TL, Kayes KJ, Feinberg J. Varus derotation osteotomy for
the treatment of hip subluxation and dislocation in cerebral palsy: statistical anal- 28. Parrott J, Boyd RN, Dobson F, Lancaster A, Love S, Oates J, Wolfe R, Nattrass
ysis in 73 hips. J Pediatr Orthop B. 2001;10:279-86. GR, Graham HK. Hip displacement in spastic cerebral palsy: repeatability of ra-
diologic measurement. J Pediatr Orthop. 2002;22:660-7.
7. Cornell MS. The hip in cerebral palsy. Dev Med Child Neurol. 1995;37:3-18.
29. Faraj S, Atherton WG, Stott NS. Inter- and intra-measurer error in the mea-
8. Herndon WA, Bolano L, Sullivan JA. Hip stabilization in severely involved cere-
surement of Reimers hip migration percentage. J Bone Joint Surg Br. 2004;
bral palsy patients. J Pediatr Orthop. 1992;12:68-73.
86:434-7.
9. de Moraes Barros Fucs PM, Svartman C, de Assumpcao RM, Kertzman PF.
30. Miller F, Bagg MR. Age and migration percentage as risk factors for progres-
Treatment of the painful chronically dislocated and subluxated hip in cerebral
sion in spastic hip disease. Dev Med Child Neurol. 1995;37:449-55.
palsy with hip arthrodesis. J Pediatr Orthop. 2003;23:529-34.
31. Miller F, Dabney KW, Rang M. Complications in cerebral palsy treatment. In:
10. Letts M, Shapiro L, Mulder K, Klassen O. The windblown hip syndrome in
Epps CH Jr, Bowen RJ, editors. Complications in pediatric orthopaedic surgery.
total body cerebral palsy. J Pediatr Orthop. 1984;4:55-62.
Philadelphia: Lippincott; 1995. p 477-544.
11. Flynn JM, Miller F. Management of hip disorders in patients with cerebral
32. Cooke PH, Cole WG, Carey RP. Dislocation of the hip in cerebral palsy. Natural
palsy. J Am Acad Orthop Surg. 2002;10:198-209.
history and predictability. J Bone Joint Surg Br. 1989;71:441-6.
12. Howard CB, McKibbin B, Williams LA, Mackie I. Factors affecting the inci-
33. Scrutton D, Baird G, Smeeton N. Hip dysplasia in bilateral cerebral palsy: inci-
dence of hip dislocation in cerebral palsy. J Bone Joint Surg Br. 1985;67:530-2.
dence and natural history in children aged 18 months to 5 years. Dev Med Child
13. Lonstein JE, Beck K. Hip dislocation and subluxation in cerebral palsy. J Pedi- Neurol. 2001;43:586-600.
atr Orthop. 1986;6:521-6.
34. Barrie JL, Galasko CS. Surgery for unstable hips in cerebral palsy. J Pediatr
14. Onimus M, Allamel G, Manzone P, Laurain JM. Prevention of hip dislocation Orthop B. 1996;5:225-31.
in cerebral palsy by early psoas and adductors tenotomies. J Pediatr Orthop.
35. Cooperman DR, Bartucci E, Dietrick E, Millar EA. Hip dislocation in spastic ce-
1991;11:432-5.
rebral palsy: long-term consequences. J Pediatr Orthop. 1987;7:268-76.
15. Scrutton D, Baird G. Surveillance measures of the hips of children with bilat-
36. Moreau M, Cook PC, Ashton B. Adductor and psoas release for subluxation of
eral cerebral palsy. Arch Dis Child. 1997;76:381-4.
the hip in children with spastic cerebral palsy. J Pediatr Orthop. 1995;15:672-6.
16. Dobson F, Boyd RN, Parrott J, Nattrass GR, Graham HK. Hip surveillance in
37. Silver RL, Rang M, Chan J, de la Garza J. Adductor release in nonambulant
children with cerebral palsy. Impact on the surgical management of spastic hip
children with cerebral palsy. J Pediatr Orthop. 1985;5:672-7.
disease. J Bone Joint Surg Br. 2002;84:720-6.
38. Winters TF Jr, Gage JR, Hicks R. Gait patterns in spastic hemiplegia in chil-
17. Kocher MS, Zurakowski D. Clinical epidemiology and biostatistics: a primer
dren and young adults. J Bone Joint Surg Am. 1987;69:437-41.
for orthopaedic surgeons. J Bone Joint Surg Am. 2004;86:607-20.
39. Beckung E, Hagberg G. Neuroimpairments, activity limitations, and participa-
18. Sanger TD, Delgado MR, Gaebler-Spira D, Hallett M, Mink JW; Task Force on
tion restrictions in children with cerebral palsy. Dev Med Child Neurol. 2002;
Childhood Motor Disorders. Classification and definition of disorders causing hy-
44:309-16.
pertonia in childhood. Pediatrics. 2003;111:e89-97.
40. Gorter JW, Rosenbaum PL, Hanna SE, Palisano RJ, Bartlett DJ, Russell DJ,
19. Sanger TD. Pathophysiology of pediatric movement disorders. J Child Neurol.
Walter SD, Raina P, Galuppi BE, Wood E. Limb distribution, motor impairment,
2003;18 Suppl 1:S9-24.
and functional classification of cerebral palsy. Dev Med Child Neurol. 2004;
20. Delgado MR, Albright AL. Movement disorders in children: definitions, classi- 46:461-7.
fications, and grading systems. J Child Neurol. 2003;18 Suppl 1:S1-8.
41. Burke RE, Fahn S, Gold AP. Delayed-onset dystonia in patients with static
21. Graham HK, Selber P. Musculoskeletal aspects of cerebral palsy. J Bone Joint encephalopathy. J Neurol Neurosurg Psychiatry. 1980;43:789-97.
Surg Br. 2003;85:157-66.
42. Saint Hilaire MH, Burke RE, Bressman SB, Brin MF, Fahn S. Delayed-onset
22. Blair E, Stanley F. Interobserver agreement in the classification of cerebral dystonia due to perinatal or early childhood asphyxia. Neurology. 1991;41:
palsy. Dev Med Child Neurol. 1985;27:615-22. 216-22.